10 Traversability

Traversability Hamiltonian Graphs Applications of Traversability
Graph Theory
Traversability
A. Senthil Thilak
Graph Theory A. Senthil Thilak

Eulerian graphs
Traversability - Eulerian graphs

A Walk is an alternating sequence of vertices and edges, starting
and ending with vertices.

Eulerian graphs

A Trail is a walk without repetition of edges. (i.e., edges are not
repeated)

Eulerian graphs

repeated)
A Circuit is a trail which starts and ends with the same vertex.

Eulerian graphs

repeated)
Definition
Euler Trail: An Euler Trail is a trail that includes all the edges.
(Remember that being a trail, it includes each edge exactly once).
A closed Euler Trail is called an Euler circuit.

Eulerian graphs

repeated)
Definition
Euler Trail: An Euler Trail is a trail that includes all the edges.
(Remember that being a trail, it includes each edge exactly once).
A closed Euler Trail is called an Euler circuit.
Definition
Euler graph/Eulerian graph: A graph G is said to be Eulerian if
it has a closed Euler trail (or) an Euler circuit.

Eulerian graphs
Example
• All cycles Cn (n ≥ 3) are Eulerian.

Eulerian graphs
Example

• Complete graph Kn is Eulerian, whenever n is odd.

Eulerian graphs
Example

• Complete graph Kn is non-Eulerian, whenever n is even.

Eulerian graphs
Example

• Complete graph Kn is non-Eulerian, whenever n is even.
• Petersen graph is non-Eulerian.
Figure: Petersen Graph

Eulerian graphs
Theorem
A graph G with δ(G ) ≥ 2, has atleast one cycle.

Eulerian graphs
Theorem
Proof.
Let G be a graph with δ(G ) ≥ 2 and P be a maximal path in G .

Eulerian graphs
Theorem
Proof.
Let u be a pendant vertex lying on P.

Eulerian graphs
Theorem
Proof.
Let u be a pendant vertex lying on P. Since P is maximal, it
cannot be extended. Hence,

Eulerian graphs
Theorem
Proof.
cannot be extended. Hence, every neighbour of u must lie on P.

Eulerian graphs
Theorem
Proof.
cannot be extended. Hence, every neighbour of u must lie on P.
Thus, ∃ a neighbour v of u such that the edge e = (u, v ) (not
included in P) together with the uv -segment of P will form a cycle.
e
u v
P

Eulerian graphs
Theorem
Proof.
Let P be a maximal path in G . Let u be a pendant vertex lying on
P. Since P is maximal, it cannot be extended. Hence, every
neighbour must already lie on P. Thus, ∃ a neighbour v of u such
that the edge e = (u, v ) (not included in P) together with the
uv -segment of P will form a cycle.
e
u v
P

Eulerian graphs
Theorem
Proof.
e
u v
P

Eulerian graphs
Theorem
Proof.
e
u v
P

Eulerian graphs
Theorem
Proof.
e
u v
P

Eulerian graphs
Theorem (Characterization of Eulerian graphs:)

A connected graph G is Eulerian if and only if all its vertices are of
even degree.

Eulerian graphs

even degree.
Proof:

Eulerian graphs

even degree.
Proof: Let G be a connected graph and suppose that G is

Eulerian.

Eulerian graphs

even degree.

Eulerian. Then it has an Euler circuit, say C , with its origin and
terminus, say u.

Eulerian graphs

even degree.

terminus, say u.
Each time a vertex v occurs as an internal vertex of C , two of the
edges incident on v are accounted for and contribute two degrees
to that vertex.

Eulerian graphs

even degree.

terminus, say u.
to that vertex.
Since an Euler circuit contains every edge of G exactly once,
deg (v ) is even, for all v 6= u.

Eulerian graphs

even degree.

terminus, say u.
to that vertex.
Similarly, C starts and ends at u and hence deg (u) is also even.

Eulerian graphs

even degree.

terminus, say u.
to that vertex.
Similarly, C starts and ends at u and hence deg (u) is also even.
Thus, all the vertices are of even degree.

Eulerian graphs
Converse part:

Eulerian graphs
Converse part: Suppose that G is connected and all its vertices

are of even degree.

Eulerian graphs

are of even degree. If possible, assume that G is non-Eulerian and
let G be the smallest non-trivial graph with this property.

Eulerian graphs

Since, degree of each vertex in G is even, δ(G ) ≥ 2 and hence,

Eulerian graphs

Since, degree of each vertex in G is even, δ(G ) ≥ 2 and hence, G
must have a circuit.

Eulerian graphs

must have a circuit. Let C be a circuit of maximum possible length
in G .

Eulerian graphs

in G . As G is non-Eulerian, C cannot include all the edges of G .

Eulerian graphs

Let G 0 be the graph obtained by removing the edges of C from G .
Then |E (G 0 )| < |E (G )|.

Eulerian graphs

Then |E (G 0 )| < |E (G )|.
Hence, by our hypothesis, G 0 must have an Euler circuit, say C 0 .

Eulerian graphs

Then |E (G 0 )| < |E (G )|.
Now, as G is connected, there must exist a vertex, say v , common
to both C and C 0 . Then CC 0 will be a circuit in G , such that
|E (CC 0 )| > |E (C )|, contradicting that

Eulerian graphs

Then |E (G 0 )| < |E (G )|.
|E (CC 0 )| > |E (C )|, contradicting that C is a ciruit of maximum
possible length in G .

Eulerian graphs

Then |E (G 0 )| < |E (G )|.
|E (CC 0 )| > |E (C )|, contradicting that C is a ciruit of maximum
possible length in G .
Hence, G must be Eulerian.

Eulerian graphs
Corollary
A connected graph G has an Euler trail if and only if it has atmost
two vertices of odd degree.
Proof.
Let G be a connected graph. Suppose G has an Euler trail, then every
vertex other than the origin and terminus are of even degree and hence,
the result follows.

Eulerian graphs
Corollary
Proof.
the result follows.
Conversely, suppose that G has atmost two (i.e., ≤ 2) vertices of odd
degree.

Eulerian graphs
Corollary
Proof.
the result follows.
degree. Suppose it has no vertex of odd degree, then it has a closed
Euler trail and the result holds.

Eulerian graphs
Corollary
Proof.
the result follows.
Euler trail and the result holds. On the other hand, if G has exatly two
vertices of odd degree, say u and v , then

Eulerian graphs
Corollary
Proof.
the result follows.
vertices of odd degree, say u and v , then obtain a new graph G 0 by
adding an edge e between u and v . (i.e, G 0 = G + e, where e = (u, v )).
Then,

Eulerian graphs
Corollary
Proof.
the result follows.
Then, every vertex in G 0 is of even degree and hence, G 0 has an Euler
circuit, say C , that includes the edge e.

Eulerian graphs
Corollary
Proof.
the result follows.
Then, every vertex in G 0 is of even degree and hence, G 0 has an Euler
circuit, say C , that includes the edge e.
Now, C − e will be an Euler trail in G .

Eulerian graphs
Exercises:
1 Which of the following figures can be drawn without lifting one’s
pen from the paper or without covering a line more than once?
2 If possible, draw a (p, q) Eulerian graph where p is even and q is

odd. Otherwise, explain why there is no such graph.
3 Show that if G has no vertex of odd degree, then G can be
partitioned into edge-disjoint cycles. (Or Show that every even
graph decomposes into cycles. [Remember that an even graph is a
graph in which all vertices are of even degree])

Eulerian graphs
Figure: Leonhard Euler (1707 - 1783) - The Father of Graph Theory

Eulerian graphs
Königsberg-Bridge problem: [Solved by Euler in 1736]

The Königsberg-Bridge problem: The city of Königsberg was
located on the Pregel river in Prussia. The city occupied two island
plus areas on both banks. These regions were linked by seven
bridges. The citizens wondered whether they could leave home,
cross every bridge exactly once and return home.
Euler proved that the problem has no solution!!!!!

Eulerian graphs
Other significant results (without proof!)

Theorem
If G is a simple graph in which every vertex has degree atleast k, then G
contains a path of length atleast k. If k ≥ 2, then G also contains a cycle
of length at least k + 1.

Eulerian graphs

Theorem
Theorem
In an even graph, every maximal trail is closed.

Eulerian graphs

Theorem
Theorem
Theorem
For a connected nontrivial graph G with exactly 2k odd vertices, where
k ≥ 1, then G can be decomposed into atleast k trails.

Eulerian graphs

Theorem
Theorem
Theorem
For a connected nontrivial graph G with exactly 2k odd vertices, where
k ≥ 1, then G can be decomposed into atleast k trails.
Making a non-Eulerian graph into Eulerian:

Hamiltonian graphs
Definition
Hamilton Path: A path that contains every vertex of G is called a
Hamilton path of G .
Hamilton Cycle: A cycle that contains every vertex of G is called
a Hamilton cycle (or spanning cycle) of G .
Definition
Hamiltonian graph: A graph G is Hamiltonian if it contains a
Hamilton cycle.


Such paths and cycles are named after Hamilton (1856), who
described, in a letter to his friend Graves, a mathematical game on
the dodecahedron in which one person sticks five pins in any five
consecutive vertices and the other is required to complete the path
so formed to a spanning cycle.
(a) Dodecahedron - (b) Herschel graph -

Hamiltonian Non-Hamiltonian

Such paths and cycles are named after Hamilton (1856), who
described, in a letter to his friend Graves, a mathematical game on
the dodecahedron in which one person sticks five pins in any five
consecutive vertices and the other is required to complete the path
so formed to a spanning cycle.
(a) Dodecahedron - (b) Herschel graph -

Hamiltonian Non-Hamiltonian
Observe that the Dodecahedron graph is 3-regular and the Herschel

graph is bipartite!!!

Examples
• The Dodecahedron given above is Hamiltonian.

• Complete graph Kn is Hamiltonian, ∀n.
• Every cycle Cn is Hamiltonian.
• The Herschel graph shown above is non-Hamiltonian.
• Petersen graph is non-Hamiltonian.

Results on Hamiltonian graphs
Theorem (A Necessary condition)

If G is Hamiltonian, then for every non-empty proper subset S of
V , ω(G − S) ≤ |S|, where ω(G − S) denotes the number of
components in G − S.
Proof.
Let G be Hamiltonian. Then,


Proof.
Let G be Hamiltonian. Then, G has a Hamilton cycle, say C .


Proof.
Let G be Hamiltonian. Then, G has a Hamilton cycle, say C . Now,
for every non-empty subset S of V , ω(C − S) ≤ |S|.


Proof.
for every non-empty subset S of V , ω(C − S) ≤ |S|. Also, C − S is
a spanning subgraph of G − S and hence,


Proof.
for every non-empty subset S of V , ω(C − S) ≤ |S|. Also, C − S is
a spanning subgraph of G − S and hence,
ω(G − S) ≤ ω(C − S) ≤ |S|. Hence, the result follows.

Sufficient condition 1 - Due to Dirac(1952)
Theorem (Sufficient condition 1 - Due to Dirac(1952))

p
If G is a simple graph with p ≥ 3 and δ(G ) ≥ , then G is
2
Hamiltonian.


p
2
Hamiltonian.
Observe that the above condition is sufficient but not necessary.


p
2
Hamiltonian.
Observe that the above condition is sufficient but not necessary.
For eg., Dodecahedron has p = 20 (≥ 3) vertices and its minimum

degree is 3 < p/2, but it is Hamiltonian.

Results on Hamiltonian graphs (contd...)

Lemma
Let G be simple and u and v be non-adjacent vertices in G such
that
deg (u) + deg (v ) ≥ p,
then G is Hamiltonian if and only if G + uv is Hamiltonian.


Lemma
that
deg (u) + deg (v ) ≥ p,
The above lemma motivates the following definition.


Lemma
that
deg (u) + deg (v ) ≥ p,
The above lemma motivates the following definition.

Definition
Closure of a graph G [C(G)]: The closure of a graph G on p
vertices, is the graph obtained from G by recursively joining pairs of
non-adjacent vertices whose degree sum is atleast p, until no such
pair remains.

Closure of a graph - Example

Example
Figure: A graph G

Example
−→

Example
−→

Example
−→

Example
−→

Example
−→

Example
−→

Example
−→

Example
−→

Example
−→

Example
−→
(a) A graph G (b) Closure of G → C (G )

Theorem
A simple graph G is Hamiltonian if and only if its closure C (G ) is
Hamiltonian.

Theorem
Hamiltonian.
Corollary
Let G be a simple graph with p ≥ 3. If C (G ) is complete, then G
is Hamiltonian.

Theorem
Hamiltonian.
Corollary
is Hamiltonian.
The above corollary can be used to deduce various sufficient
conditions for a graph to be Hamiltonian in terms of its vertex
degrees.

Theorem
Hamiltonian.
Corollary
is Hamiltonian.
degrees.
p
For example, when δ(G ) ≥ , clearly C (G ) is always complete.
2

Theorem
Hamiltonian.
Corollary
is Hamiltonian.
degrees.
p
For example, when δ(G ) ≥ , clearly C (G ) is always complete. So,
2
Dirac’s condition follows as an immediate corollary.

Theorem
Hamiltonian.
Corollary
is Hamiltonian.
degrees.
p
For example, when δ(G ) ≥ , clearly C (G ) is always complete. So,
2
Dirac’s condition follows as an immediate corollary. A more general
condition than that of Dirac was obtained by Chvátal (1972), which
is stated below.
Sufficient condition 2 - Due to Chvátal (1974)
Theorem (Sufficient condition 2 - Due to Chvátal (1974) -

Stronger than Dirac’s)
Let G be a simple graph with degree sequence (d1 , d2 , . . . dp ),
where d1 ≤ d2 ≤ · · · ≤ dp with p ≥ 3. Suppose that there is no
value of m < p/2 for which dm ≤ m and dp−m < p − m, then G is
Hamiltonian.

The Chinese Postman Problem

The Chinese Postman Problem: In his job, a postman picks up
mail at the post office, delivers it and then returns to the post
office.


office. He must, of course, cover each street in his area at least
once. Subject to this condition, he wishes to choose his route in
such a way that he walks as little as possible. This problem is
known as the the Chinese postman problem, since it was first
considered by a Chinese Mathematician, Kuan (1962).


office. He must, of course, cover each street in his area at least
once. Subject to this condition, he wishes to choose his route in
such a way that he walks as little as possible. This problem is
known as the the Chinese postman problem, since it was first
considered by a Chinese Mathematician, Kuan (1962).
The problem is modelled as a weighted graph, which is defined as a
graph in which real values are assigned to each edge, referred to as
the weight of that edge.
The weight of a tour/circuit v0 e1 v1 . . . en v0 is defined as
Xn
w (ei ).
i=1

Clearly, the Chinese postman problem is just that of finding a

minimum-weight tour in a weighted connected graph with
non-negative weights. (Such a tour is generally referred to as an
optimal tour.)


optimal tour.)
Note: If G is Eulerian, then any Euler tour of G is an optimal
tour, because an Euler tour is a tour that traverses each
edge exactly once.


optimal tour.)
Note: If G is Eulerian, then any Euler tour of G is an optimal
tour, because an Euler tour is a tour that traverses each
edge exactly once.
Thus, the Chinese postman problem is easily solved in this case,
since there exists a good algorithm for determining an Euler tour in
an Eulerian graph. Fleury’s Algorithm is one such algorithm!!!

Travelling salesman Problem
The Travelling Salesman Problem: A travelling salesman wishes to

visit a number of towns and then return to his starting point.


visit a number of towns and then return to his starting point. Given the
travelling times between towns, how should he plan his travel so that he
visits each town exactly once and travels in all with as short time as
possible? This problem is known as the Travelling Salesman Problem.


The problem is again modelled as a weighted graph and the aim is to find
a minimum weighted Hamilton cycle (or an optimal cycle) in a weighted
complete graph.


The problem is again modelled as a weighted graph and the aim is to find
a minimum weighted Hamilton cycle (or an optimal cycle) in a weighted
complete graph.
In contrast with the shortest path problem and the connector problem, no
efficient algorithm (global solution)for solving the travelling salesman
problem is known. But, several good approximation procedures (local
solutions) have been attempted. There are also necessary and sufficient
conditions to determine if a possible solution does exist when a given
graph is not complete.

10 Traversability

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

10 Traversability

Uploaded by

Copyright:

Available Formats

Traversability Hamiltonian Graphs Applications of Traversability

Graph Theory A. Senthil Thilak

Traversability - Eulerian graphs

Graph Theory A. Senthil Thilak

Traversability - Eulerian graphs

Graph Theory A. Senthil Thilak

Traversability - Eulerian graphs

Graph Theory A. Senthil Thilak

Traversability - Eulerian graphs

Graph Theory A. Senthil Thilak

Traversability - Eulerian graphs

Graph Theory A. Senthil Thilak

• All cycles Cn (n ≥ 3) are Eulerian.

Graph Theory A. Senthil Thilak

• All cycles Cn (n ≥ 3) are Eulerian.

Graph Theory A. Senthil Thilak

• All cycles Cn (n ≥ 3) are Eulerian.

Graph Theory A. Senthil Thilak

• All cycles Cn (n ≥ 3) are Eulerian.

Figure: Petersen Graph

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Theorem (Characterization of Eulerian graphs:)

Graph Theory A. Senthil Thilak

Theorem (Characterization of Eulerian graphs:)

Graph Theory A. Senthil Thilak

Theorem (Characterization of Eulerian graphs:)

Proof: Let G be a connected graph and suppose that G is

Graph Theory A. Senthil Thilak

Theorem (Characterization of Eulerian graphs:)

Proof: Let G be a connected graph and suppose that G is

Graph Theory A. Senthil Thilak

Theorem (Characterization of Eulerian graphs:)

Proof: Let G be a connected graph and suppose that G is

Graph Theory A. Senthil Thilak

Theorem (Characterization of Eulerian graphs:)

Proof: Let G be a connected graph and suppose that G is

Graph Theory A. Senthil Thilak

Theorem (Characterization of Eulerian graphs:)

Proof: Let G be a connected graph and suppose that G is

Graph Theory A. Senthil Thilak

Theorem (Characterization of Eulerian graphs:)

Proof: Let G be a connected graph and suppose that G is

Graph Theory A. Senthil Thilak

Graph Theory A. Senthil Thilak

Converse part: Suppose that G is connected and all its vertices

Graph Theory A. Senthil Thilak

Converse part: Suppose that G is connected and all its vertices

Graph Theory A. Senthil Thilak

Converse part: Suppose that G is connected and all its vertices

Graph Theory A. Senthil Thilak

Converse part: Suppose that G is connected and all its vertices

Graph Theory A. Senthil Thilak

Converse part: Suppose that G is connected and all its vertices